Methods to control for confounding - Introduction & Overview - Nicolle M Gatto 18 February 2015
|
|
- Morgan Blair
- 6 years ago
- Views:
Transcription
1 Methods to control for confounding - Introduction & Overview - Nicolle M Gatto 18 February 2015
2 Learning Objectives At the end of this confounding control overview, you will be able to: Understand how confounding arises in pharmacoepidemiology and the problem it poses for causal inference Differentiate (exposure probability) weighting approaches to confounding control, including propensity score control, matching, and stratification, and inverse probability weighting Relate these weighting approaches to randomization List key assumptions necessary for these approaches to estimate causal effects Distinguish between point treatment and time-dependent confounding 2
3 Outline Ideal comparisons for causal inference How confounding impedes causal inference The gold standard Mimicking ideal comparisons without randomization: Typical multivariate regression Propensity score methods Inverse probability weighting (IPW) Practical considerations Take home messages 3
4 Aim of comparative pharmacoepidemiology studies Does drug A protect against outcome Z? A? Z Does drug A cause outcome Y? A? Y Here, the exposure of interest, A, is a point-treatment 4
5 The essential problem L1 L2 L i Drug A Outside the context of study, many reasons people take one medication over another 5
6 The essential problem L1 L2 L i Drug A? Adverse effect Y These reasons are often related to outcomes of interest 6
7 The essential problem L1 L2 L i Drug A? Adverse effect Y Confounding: When risk factors for the disease are associated with use of the medication of interest 7
8 The essential problem L Drug A? Adverse effect Y For simplicity let s consider L as the totality of confounding Point-treatment confounding 8
9 The ideal (hint: it s not the RCT) Target Population N N Y+ if A+ (Y+ if A+ / N) N Y+ if A- (Y+ if A- / N) A type of counterfactual comparison = RD causal 9
10 The ideal (hint: it s not the RCT) We ve defined RD causal (for now) as the effect with all else equal - in the total population if everyone was exposed versus no one was exposed But at least one of these conditions is counter-tofact Instead we use substitutes for the ideal risks 10
11 The real world substitute In target Population N, people become A+ or A- N f1 1-f1 A+ Y+ in A+ (Y+ in A+ / A+) A- Y+ in A- (Y+ in A- / A-) An observable estimate = RD crude 11
12 The real world substitute In target Population N, people become A+ or A- N f1 1-f1 A+ Y+ in A+ (Y+ in A+ / [f 1 N]) A- Y+ in A- (Y+ in A- / A-) An observable estimate = RD crude 12
13 The real world substitute In target Population N, people become A+ or A- N f1 A+ Y+ in A+ (Y+ in A+ / [f 1 N]) 1-f1 A- Y+ in A- (Y+ in A- / [(1-f 1 )N]) An observable estimate = RD crude 13
14 The real world substitute In target Population N, people become A+ or A- N f1 A+ Y+ in A+ ([f 2 f 1 Y+ if A+ ] / [f 1 N]) 1-f1 A- Y+ in A- (Y+ in A- / [(1-f 1 )N]) An observable estimate = RD crude 14
15 The real world substitute In target Population N, people become A+ or A- N f1 1-f1 A+ Y+ in A+ ([f 2 f 1 Y+ if A+ ] / [f 1 N]) A- Y+ in A- ([f 3 (1-f 1 )Y+ if A- ]/ [(1-f 1 )N]) An observable estimate = RD crude 15
16 The real world substitute In target Population N, people become A+ or A- N f1 1-f1 A+ Y+ in A+ ([f 2 f 1 Y+ if A+ ] / [f 1 N]) A- Y+ in A- ([f 3 (1-f 1 )Y+ if A- ]/ [(1-f 1 )N]) With confounding by L RD causal RD crude 16
17 Possible solutions Try to avoid confounding through design, e.g. Randomize Match Restrict Try to control confounding through analysis, e.g.: Traditional analysis methods Techniques that weight by causes of the treatment (aka mimic randomization) Techniques that weight by causes of the outcome 17
18 Why randomization is the gold standard Target Population N, randomized 1:1 to A+ or A- N f1 1-f1 A+ Y+ in A+ (Y+ in A+ / A+) A- Y+ in A- (Y+ in A- / A-) An observable estimate = RD crude 18
19 Why randomization is the gold standard Target Population N, randomized 1:1 to A+ or A- N f1 1-f1 A+ Y+ in A+ (Y+ in A+ / 0.5N) A- Y+ in A- (Y+ in A- / A-) An observable estimate = RD crude 19
20 Why randomization is the gold standard Target Population N, randomized 1:1 to A+ or A- N f1 1-f1 A+ Y+ in A+ (0.5Y+ if A+ / 0.5N) A- Y+ in A- (Y+ in A- / A-) An observable estimate = RD crude 20
21 Why randomization is the gold standard Target Population N, randomized 1:1 to A+ or A- N f1 1-f1 A+ Y+ in A+ (0.5Y+ if A+ / 0.5N) A- Y+ in A- (0.5Y+ if A- / 0.5N) An observable estimate RD causal = RD crude 21
22 Why randomization is the gold standard Target Population N, randomized 1:3 to A+ or A- N f1 1-f1 A+ Y+ in A+ (0.25Y+ if A+ / 0.25N) A- Y+ in A- (0.75Y+ if A- / 0.75N) RD causal = RD crude 22
23 Why randomization is the gold standard L Drug A? Adverse effect Y If randomization is successful, no confounding at baseline 23
24 Possible solutions Try to avoid confounding through design, e.g. Randomize Match Restrict Try to control confounding through analysis, e.g.: Traditional analysis methods Techniques that weight by causes of the treatment (aka mimic randomization) Techniques that weight by causes of the outcome 24
25 Basic requirements of these methods Appropriately hypothesize factors that contribute to confounding Can measure these factors (validly)... there is no unmeasured confounding 25
26 Option 1: Traditional confounding control For example, multivariate regression analysis Estimates RD causal? Yes, when: All variables that contribute to confounding are included in the model Statistical model is specified correctly Effect of A Y across strata of confounding variables is homogeneous 26
27 Option 2: Propensity score (PS) methods Estimates RD causal? Yes, when: All variables that contribute to confounding are included in the model Statistical model is specified correctly Effect of A Y across strata of confounding variables is homogeneous (to estimate RD causal as we ve defined it: the total population effect) 27
28 Option 3: Inverse probability weighting (IPW) Estimates RD causal? Yes, when: All variables that contribute to confounding are included in the model Statistical model is specified correctly Positivity holds There are no strata of L for which there are no subjects who received a particular exposure level 28
29 To estimate RD causal in total population Characteristic Requires all confounding factors measured? Statistical correctly specified Can handle large number of confounding variables? Traditional methods Propensity score control IPW Yes Yes Yes Yes Yes Yes No Yes Yes Requires homogeneity? Yes Yes No Requires positivity? No No Yes 29
30 Estimating the propensity score Traditional PS: probability of being exposed conditional on predictors of exposure Minimally sufficient PS: probability of being exposed conditional on variables that contribute to confounding 30
31 Estimating the propensity score Traditional PS: probability of being exposed conditional on predictors of exposure Minimally sufficient PS: probability of being exposed conditional on variables that contribute to confounding A weight is assigned to each person according to the conditional probability of exposure given the confounder(s) 31
32 Estimating the propensity score: Example 1 Study Population (N) L+ L- A+ A- A+ A- Y+ Y- Y+ Y- Y+ Y- Y+ Y- Weight these people by pr(a+ L+) 32
33 Estimating the propensity score: Example 2 Study Population (N) L+ L- A+ A- A+ A- Y+ Y- Y+ Y- Y+ Y- Y+ Y- Weight these people by pr(a+ L-) 33
34 Estimating the propensity score Example SAS code: **estimate propensity scores**; proc logistic descending data=dataset1; model A=L1 L2 L4 L5 L7 L8; output out=proba1_ps p=pra1_ps xbeta=logit_ps; run; 34
35 Applying the propensity score Options: PS control, stratification & matching After estimating PS for each person, use it to control confounding by: Adjusting for the PS as a covariate Stratifying the effect estimate by the PS Matching A+ and A- people by PS 35
36 Applying the propensity score Example SAS code: **PS adjusted (as covariate) RR**; proc genmod data=proba1_ps; model Y=A pra1_ps / dist=poisson; run; 36
37 Applying the propensity score Example SAS code: **PS quintile stratification for CMH RR**; proc rank data=proba1_ps groups=5 out=tempps; ranks rnks; var pra1_ps; run; proc sort data=tempps; by A Y; proc freq data=tempps order=data; tables rnks*a*y / cmh; output out=tempps1 mhrrc2 lgrrc2; run; data PSstrat_5; set tempps1; beta=log(1/_mhrrc2_); SE=(log(U_MHRRC2)-log(L_MHRRC2))/3.92; run; 37
38 Estimating the inverse probability weight Another way to use the propensity score! IPW: Inverse of the exposure probability conditional on variables that contribute to confounding A weight is assigned to each person according to the inverse of the conditional probability of his/her exposure status given the confounder(s) 38
39 Estimating the IPW: Example 1 Study Population (N) L+ L- A+ A- A+ A- Y+ Y- Y+ Y- Y+ Y- Y+ Y- Weight these people by 1/ pr(a- L+) 39
40 Estimating the IPW: Example 2 Study Population (N) L+ L- A+ A- A+ A- Y+ Y- Y+ Y- Y+ Y- Y+ Y- Weight these people by 1 / pr(a+ L-) 40
41 Estimating the IPW Example SAS code: **estimate propensity scores**; proc logistic descending data=dataset1; model A=L1 L2 L4 L5 L7 L8; output out=proba1_ps p=pra1_ps xbeta=logit_ps; run; data weights_ps; set proba1_ps; pra0_ps=1-pra1_ps; if A=1 then IPW=1/prA1_PS; else IPW=1/prA0_PS; run; 41
42 Applying the IPW Example SAS code: **IP weighted RR in total population**; proc genmod data=weights_ps; class ID; model Y=A /dist=poisson; weight IPW; repeated subject=id; estimate 'beta' A 1; run; 42
43 How do these methods relate to RCTs? They use weighting to mimic randomization PS methods estimate effect using weighting across x number of new populations IPW weights people by the IPW, creating a pseudopopulation 43
44 The pseudo-population Study Population (N) L+ L- A+ A- A+ A- Y+ Y- Y+ Y- Y+ Y- Y+ Y- 1/ pr(a+ L+) 1/ pr(a- L+) 1/ pr(a+ L-) 1/ pr(a- L-) 44
45 The pseudo-population Study Population (N) L+ L- A+ A- A+ A- Y+ Y- Y+ Y- Y+ Y- Y+ Y- 1/ 0.9 = 1.1 1/ 0.1 = / 0.05 = / 0.95 =
46 Some practical considerations 46
47 Propensity score balancing Untreated Treated Propensity Score 47
48 Propensity score balancing Untreated Treated Propensity Score 48
49 How will PS be estimated? For example: Covariates to be included Interaction terms to be included Type of model used to estimate the probability of exposure Type of regression used to model the outcome 49
50 How will IPW be estimated? For example: Covariates to be included Interaction terms to be included Type of model used to estimate the probability of exposure Type of regression used to model the outcome Stabilized or unstabilized weights Robust variance estimation 50
51 To estimate RD causal in total population Characteristic Requires all confounding factors measured? Statistical correctly specified Can handle large number of confounding variables? Traditional methods Propensity score control IPW Yes Yes Yes Yes Yes Yes No Yes Yes Requires homogeneity? Yes Yes No Requires positivity? No No Yes 51
52 To estimate RD causal in treated Characteristic Requires all confounding factors measured? Statistical correctly specified Can handle large number of confounding variables? Traditional methods Propensity score control IPW Yes Yes Yes Yes Yes Yes No Yes Yes Requires homogeneity? Yes No* No # Requires positivity? No No Yes * If PS matching or SMR weighting to combine PS strata # If SMR weighting used to create pseudo-population 52
53 Instrumental variable analysis: A solution? Stay tuned... 53
54 What if we are interested in a time varying treatment effect? A 0 A 1 Y Treatment (A) varies with time 54
55 What if we are interested in a time varying treatment effect? A 0 A 1 Y Treatment (A) varies with time the causal effect if everyone in the population had been A ++ versus if everyone had been A -- 55
56 What if we are interested in a time varying treatment effect? L 0 L 1 A 0 A 1 Y Treatment (A) varies with time Covariate (L) varies with time 56
57 What if we are interested in a time varying treatment effect? L 0 L 1 A 0 A 1 Y Treatment (A) varies with time Covariate (L) varies with time and: Predicts the outcome 57
58 What if we are interested in a time varying treatment effect? L 0 L 1 A 0 A 1 Y Treatment (A) varies with time Covariate (L) varies with time and: Predicts the outcome Predicts exposure 58
59 What if we are interested in a time varying treatment effect? L 0 L 1 A 0 A 1 Y Treatment (A) varies with time Covariate (L) varies with time and: Predicts the outcome Predicts exposure Is affected by previous exposure How do we address the confounding and mediational effects of L? 59
60 Inverse probability weighting: A solution? Stay tuned... 60
61 Take home messages These weighting approaches require all confounding factors are hypothesized and measured To estimate the causal effect in the total population, only IPW does not require homogeneity of the treatment effect The choice of confounding control method for pointtreatment confounding should be driven by: Causal question / target population Practical considerations 61
62 References Hernan MA and Robins JM. Estimating causal effects from epidemiological data. J Epidemiol Community Health 2006;60: Hernan MA et al. Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. Epidemiology 2000;11: Robins JM et al. Marginal structural models and causal inference in epidemiology. Epidemiology 2000;11: Sato T and Matsuyama Y. Marginal Structural Models as a Tool for Standardization. Epidemiology 2003;14:680-6 Sturmer T, Rothman KJ, Glynn RJ Insights into different results from different causal contrasts in the presence of effect-measure modification. Pharmacoepidemiol Drug Saf 2006; 15,
63 Thank you 63
Propensity Score Methods to Adjust for Bias in Observational Data SAS HEALTH USERS GROUP APRIL 6, 2018
Propensity Score Methods to Adjust for Bias in Observational Data SAS HEALTH USERS GROUP APRIL 6, 2018 Institute Institute for Clinical for Clinical Evaluative Evaluative Sciences Sciences Overview 1.
More informationInsights into different results from different causal contrasts in the presence of effect-measure modification y
pharmacoepidemiology and drug safety 2006; 15: 698 709 Published online 10 March 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/pds.1231 ORIGINAL REPORT Insights into different results
More informationPropensity scores for confounder adjustment when assessing the effects of medical interventions using nonexperimental study designs
Review Propensity scores for confounder adjustment when assessing the effects of medical interventions using nonexperimental study designs T. St urmer 1, R. Wyss 1, R. J. Glynn 2 & M. A. Brookhart 1 doi:
More informationAssessing the impact of unmeasured confounding: confounding functions for causal inference
Assessing the impact of unmeasured confounding: confounding functions for causal inference Jessica Kasza jessica.kasza@monash.edu Department of Epidemiology and Preventive Medicine, Monash University Victorian
More informationInverse Probability of Censoring Weighting for Selective Crossover in Oncology Clinical Trials.
Paper SP02 Inverse Probability of Censoring Weighting for Selective Crossover in Oncology Clinical Trials. José Luis Jiménez-Moro (PharmaMar, Madrid, Spain) Javier Gómez (PharmaMar, Madrid, Spain) ABSTRACT
More informationApplication of Propensity Score Models in Observational Studies
Paper 2522-2018 Application of Propensity Score Models in Observational Studies Nikki Carroll, Kaiser Permanente Colorado ABSTRACT Treatment effects from observational studies may be biased as patients
More informationManitoba Centre for Health Policy. Inverse Propensity Score Weights or IPTWs
Manitoba Centre for Health Policy Inverse Propensity Score Weights or IPTWs 1 Describe Different Types of Treatment Effects Average Treatment Effect Average Treatment Effect among Treated Average Treatment
More informationIntroducing a SAS macro for doubly robust estimation
Introducing a SAS macro for doubly robust estimation 1Michele Jonsson Funk, PhD, 1Daniel Westreich MSPH, 2Marie Davidian PhD, 3Chris Weisen PhD 1Department of Epidemiology and 3Odum Institute for Research
More informationThe Logic of Causal Inference. Uwe Siebert, MD, MPH, MSc, ScD
The Logic of Causal Inference Uwe Siebert, MD, MPH, MSc, ScD (1) Dept. of Public Health, Health Services Research and Health Technology Assessment, UMIT - University for Health Sciences, Medical Informatics
More informationCAUSAL INFERENCE IN HIV/AIDS RESEARCH: GENERALIZABILITY AND APPLICATIONS
CAUSAL INFERENCE IN HIV/AIDS RESEARCH: GENERALIZABILITY AND APPLICATIONS Ashley L. Buchanan A dissertation submitted to the faculty at the University of North Carolina at Chapel Hill in partial fulfillment
More informationPropensity score methods to adjust for confounding in assessing treatment effects: bias and precision
ISPUB.COM The Internet Journal of Epidemiology Volume 7 Number 2 Propensity score methods to adjust for confounding in assessing treatment effects: bias and precision Z Wang Abstract There is an increasing
More informationPubH 7405: REGRESSION ANALYSIS. Propensity Score
PubH 7405: REGRESSION ANALYSIS Propensity Score INTRODUCTION: There is a growing interest in using observational (or nonrandomized) studies to estimate the effects of treatments on outcomes. In observational
More informationDeclaration of interests. Register-based research on safety and effectiveness opportunities and challenges 08/04/2018
Register-based research on safety and effectiveness opportunities and challenges Morten Andersen Department of Drug Design And Pharmacology Declaration of interests Participate(d) in research projects
More informationComparisons of Dynamic Treatment Regimes using Observational Data
Comparisons of Dynamic Treatment Regimes using Observational Data Bryan Blette University of North Carolina at Chapel Hill 4/19/18 Blette (UNC) BIOS 740 Final Presentation 4/19/18 1 / 15 Overview 1 Motivation
More informationTime-varying confounding and marginal structural model
Time-varying confounding and marginal structural model By David Chyou Overview An overview of time-varying confounding. Marginal structural model. Weighing regression by propensity score. Future prospective.
More informationBIOSTATISTICAL METHODS
BIOSTATISTICAL METHODS FOR TRANSLATIONAL & CLINICAL RESEARCH PROPENSITY SCORE Confounding Definition: A situation in which the effect or association between an exposure (a predictor or risk factor) and
More informationComparison And Application Of Methods To Address Confounding By Indication In Non- Randomized Clinical Studies
University of Massachusetts Amherst ScholarWorks@UMass Amherst Masters Theses 1911 - February 2014 Dissertations and Theses 2013 Comparison And Application Of Methods To Address Confounding By Indication
More informationEvaluating health management programmes over time: application of propensity score-based weighting to longitudinal datajep_
Journal of Evaluation in Clinical Practice ISSN 1356-1294 Evaluating health management programmes over time: application of propensity score-based weighting to longitudinal datajep_1361 180..185 Ariel
More informationBy: Mei-Jie Zhang, Ph.D.
Propensity Scores By: Mei-Jie Zhang, Ph.D. Medical College of Wisconsin, Division of Biostatistics Friday, March 29, 2013 12:00-1:00 pm The Medical College of Wisconsin is accredited by the Accreditation
More informationBrief introduction to instrumental variables. IV Workshop, Bristol, Miguel A. Hernán Department of Epidemiology Harvard School of Public Health
Brief introduction to instrumental variables IV Workshop, Bristol, 2008 Miguel A. Hernán Department of Epidemiology Harvard School of Public Health Goal: To consistently estimate the average causal effect
More informationWorth the weight: Using Inverse Probability Weighted Cox Models in AIDS Research
University of Rhode Island DigitalCommons@URI Pharmacy Practice Faculty Publications Pharmacy Practice 2014 Worth the weight: Using Inverse Probability Weighted Cox Models in AIDS Research Ashley L. Buchanan
More informationChapter 10. Considerations for Statistical Analysis
Chapter 10. Considerations for Statistical Analysis Patrick G. Arbogast, Ph.D. (deceased) Kaiser Permanente Northwest, Portland, OR Tyler J. VanderWeele, Ph.D. Harvard School of Public Health, Boston,
More informationHow should the propensity score be estimated when some confounders are partially observed?
How should the propensity score be estimated when some confounders are partially observed? Clémence Leyrat 1, James Carpenter 1,2, Elizabeth Williamson 1,3, Helen Blake 1 1 Department of Medical statistics,
More informationMatt Laidler, MPH, MA Acute and Communicable Disease Program Oregon Health Authority. SOSUG, April 17, 2014
Matt Laidler, MPH, MA Acute and Communicable Disease Program Oregon Health Authority SOSUG, April 17, 2014 The conditional probability of being assigned to a particular treatment given a vector of observed
More informationPHA 6935 ADVANCED PHARMACOEPIDEMIOLOGY
PHA 6935 ADVANCED PHARMACOEPIDEMIOLOGY Course Description: PHA6935 is a graduate level course that is structured as an interactive discussion of selected readings based on topics in contemporary pharmacoepidemiology
More informationSection 1: The Science of Clinical Investigation
INTRODUCTION TO CLINICAL RESEARCH Scientific Concepts for Clinical Research Scott L. Zeger, Ph.D. July 14, 2014 Section 1: The Science of Clinical Investigation 1. Platonic model: science as the search
More informationEstimating Direct Effects of New HIV Prevention Methods. Focus: the MIRA Trial
Estimating Direct Effects of New HIV Prevention Methods. Focus: the MIRA Trial Michael Rosenblum UCSF: Helen Cheng, Nancy Padian, Steve Shiboski, Ariane van der Straten Berkeley: Nicholas P. Jewell, Mark
More informationImproved control for confounding using propensity scores and instrumental variables?
Improved control for confounding using propensity scores and instrumental variables? Dr. Olaf H.Klungel Dept. of Pharmacoepidemiology & Clinical Pharmacology, Utrecht Institute of Pharmaceutical Sciences
More informationApplication of Marginal Structural Models in Pharmacoepidemiologic Studies
Virginia Commonwealth University VCU Scholars Compass Theses and Dissertations Graduate School 2014 Application of Marginal Structural Models in Pharmacoepidemiologic Studies Shibing Yang Virginia Commonwealth
More informationSupplement 2. Use of Directed Acyclic Graphs (DAGs)
Supplement 2. Use of Directed Acyclic Graphs (DAGs) Abstract This supplement describes how counterfactual theory is used to define causal effects and the conditions in which observed data can be used to
More informationPropensity Score Methods for Estimating Causality in the Absence of Random Assignment: Applications for Child Care Policy Research
2012 CCPRC Meeting Methodology Presession Workshop October 23, 2012, 2:00-5:00 p.m. Propensity Score Methods for Estimating Causality in the Absence of Random Assignment: Applications for Child Care Policy
More informationPropensity Score Methods for Causal Inference with the PSMATCH Procedure
Paper SAS332-2017 Propensity Score Methods for Causal Inference with the PSMATCH Procedure Yang Yuan, Yiu-Fai Yung, and Maura Stokes, SAS Institute Inc. Abstract In a randomized study, subjects are randomly
More informationNIH Public Access Author Manuscript Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2014 May 23.
NIH Public Access Author Manuscript Published in final edited form as: Circ Cardiovasc Qual Outcomes. 2013 September 1; 6(5): 604 611. doi:10.1161/circoutcomes. 113.000359. Propensity Score Methods for
More informationPropensity scores: what, why and why not?
Propensity scores: what, why and why not? Rhian Daniel, Cardiff University @statnav Joint workshop S3RI & Wessex Institute University of Southampton, 22nd March 2018 Rhian Daniel @statnav/propensity scores:
More informationHandling time varying confounding in observational research
Handling time varying confounding in observational research Mohammad Ali Mansournia, 1 Mahyar Etminan, 2 Goodarz Danaei, 3 Jay S Kaufman, 4 Gary Collins 5 1 Department of Epidemiology and Biostatistics,
More informationPropensity score analysis with the latest SAS/STAT procedures PSMATCH and CAUSALTRT
Propensity score analysis with the latest SAS/STAT procedures PSMATCH and CAUSALTRT Yuriy Chechulin, Statistician Modeling, Advanced Analytics What is SAS Global Forum One of the largest global analytical
More informationMoving beyond regression toward causality:
Moving beyond regression toward causality: INTRODUCING ADVANCED STATISTICAL METHODS TO ADVANCE SEXUAL VIOLENCE RESEARCH Regine Haardörfer, Ph.D. Emory University rhaardo@emory.edu OR Regine.Haardoerfer@Emory.edu
More informationRecent developments for combining evidence within evidence streams: bias-adjusted meta-analysis
EFSA/EBTC Colloquium, 25 October 2017 Recent developments for combining evidence within evidence streams: bias-adjusted meta-analysis Julian Higgins University of Bristol 1 Introduction to concepts Standard
More informationJoseph W Hogan Brown University & AMPATH February 16, 2010
Joseph W Hogan Brown University & AMPATH February 16, 2010 Drinking and lung cancer Gender bias and graduate admissions AMPATH nutrition study Stratification and regression drinking and lung cancer graduate
More informationMethodology for Non-Randomized Clinical Trials: Propensity Score Analysis Dan Conroy, Ph.D., inventiv Health, Burlington, MA
PharmaSUG 2014 - Paper SP08 Methodology for Non-Randomized Clinical Trials: Propensity Score Analysis Dan Conroy, Ph.D., inventiv Health, Burlington, MA ABSTRACT Randomized clinical trials serve as the
More informationObservational comparative effectiveness research using large databases
Observational comparative effectiveness research using large databases Robert W. Yeh, MD MSc Richard A. and Susan F. Smith Center for Outcomes Research in Cardiology Beth Israel Deaconess Medical Center
More informationOHDSI Tutorial: Design and implementation of a comparative cohort study in observational healthcare data
OHDSI Tutorial: Design and implementation of a comparative cohort study in observational healthcare data Faculty: Martijn Schuemie (Janssen Research and Development) Marc Suchard (UCLA) Patrick Ryan (Janssen
More informationEpidemiologic Methods I & II Epidem 201AB Winter & Spring 2002
DETAILED COURSE OUTLINE Epidemiologic Methods I & II Epidem 201AB Winter & Spring 2002 Hal Morgenstern, Ph.D. Department of Epidemiology UCLA School of Public Health Page 1 I. THE NATURE OF EPIDEMIOLOGIC
More informationBeyond Controlling for Confounding: Design Strategies to Avoid Selection Bias and Improve Efficiency in Observational Studies
September 27, 2018 Beyond Controlling for Confounding: Design Strategies to Avoid Selection Bias and Improve Efficiency in Observational Studies A Case Study of Screening Colonoscopy Our Team Xabier Garcia
More informationConfounding and Bias
28 th International Conference on Pharmacoepidemiology and Therapeutic Risk Management Barcelona, Spain August 22, 2012 Confounding and Bias Tobias Gerhard, PhD Assistant Professor, Ernest Mario School
More informationPropensity scores and instrumental variables to control for confounding. ISPE mid-year meeting München, 2013 Rolf H.H. Groenwold, MD, PhD
Propensity scores and instrumental variables to control for confounding ISPE mid-year meeting München, 2013 Rolf H.H. Groenwold, MD, PhD WP2 WG2: aims Evaluate methods to control for observed and unobserved
More informationA Practical Guide to Getting Started with Propensity Scores
Paper 689-2017 A Practical Guide to Getting Started with Propensity Scores Thomas Gant, Keith Crowland Data & Information Management Enhancement (DIME) Kaiser Permanente ABSTRACT This paper gives tools
More informationAn Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies
Multivariate Behavioral Research, 46:399 424, 2011 Copyright Taylor & Francis Group, LLC ISSN: 0027-3171 print/1532-7906 online DOI: 10.1080/00273171.2011.568786 An Introduction to Propensity Score Methods
More informationin Longitudinal Studies
Balancing Treatment Comparisons in Longitudinal Studies Sue M. Marcus, PhD, is with the Department of Psychiatry, Mount Sinai School of Medicine, New York. Juned Siddique, DrPH, is with the Department
More informationINTRODUCTION TO SURVIVAL CURVES
SURVIVAL CURVES WITH NON-RANDOMIZED DESIGNS: HOW TO ADDRESS POTENTIAL BIAS AND INTERPRET ADJUSTED SURVIVAL CURVES Workshop W25, Wednesday, May 25, 2016 ISPOR 21 st International Meeting, Washington, DC,
More informationRise of the Machines
Rise of the Machines Statistical machine learning for observational studies: confounding adjustment and subgroup identification Armand Chouzy, ETH (summer intern) Jason Wang, Celgene PSI conference 2018
More informationLecture Outline. Biost 517 Applied Biostatistics I. Purpose of Descriptive Statistics. Purpose of Descriptive Statistics
Biost 517 Applied Biostatistics I Scott S. Emerson, M.D., Ph.D. Professor of Biostatistics University of Washington Lecture 3: Overview of Descriptive Statistics October 3, 2005 Lecture Outline Purpose
More informationPitfalls to Avoid in Observational Studies in the ICU
Pitfalls to Avoid in Observational Studies in the ICU Rob Fowler, MDCM, MS(Epi), FRCP Department of Medicine & Critical Care Medicine Sunnybrook Hospital Institute of Health Policy, Management and Evaluation
More informationBy: Armend Lokku Supervisor: Dr. Lucia Mirea. Maternal-Infant Care Research Center, Mount Sinai Hospital
By: Armend Lokku Supervisor: Dr. Lucia Mirea Maternal-Infant Care Research Center, Mount Sinai Hospital Background My practicum placement was at the Maternal-Infant Care Research Center (MiCare) at Mount
More informationUsing Propensity Score Matching in Clinical Investigations: A Discussion and Illustration
208 International Journal of Statistics in Medical Research, 2015, 4, 208-216 Using Propensity Score Matching in Clinical Investigations: A Discussion and Illustration Carrie Hosman 1,* and Hitinder S.
More informationGeorge B. Ploubidis. The role of sensitivity analysis in the estimation of causal pathways from observational data. Improving health worldwide
George B. Ploubidis The role of sensitivity analysis in the estimation of causal pathways from observational data Improving health worldwide www.lshtm.ac.uk Outline Sensitivity analysis Causal Mediation
More informationCAN EFFECTIVENESS BE MEASURED OUTSIDE A CLINICAL TRIAL?
CAN EFFECTIVENESS BE MEASURED OUTSIDE A CLINICAL TRIAL? Mette Nørgaard, Professor, MD, PhD Department of Clinical Epidemiology Aarhus Universitety Hospital Aarhus, Denmark Danish Medical Birth Registry
More informationI. Identifying the question Define Research Hypothesis and Questions
Term Paper I. Identifying the question What is the question? (What are my hypotheses?) Is it possible to answer the question with statistics? Is the data obtainable? (birth weight, socio economic, drugs,
More informationAccounting for treatment switching/discontinuation in comparative effectiveness studies
Global Medical Affairs RWE and Digital Accounting for treatment switching/discontinuation in comparative effectiveness studies Melvin Skip Olson, PhD Global Head, Real World Data Strategy and Innovation
More informationThe ROBINS-I tool is reproduced from riskofbias.info with the permission of the authors. The tool should not be modified for use.
Table A. The Risk Of Bias In Non-romized Studies of Interventions (ROBINS-I) I) assessment tool The ROBINS-I tool is reproduced from riskofbias.info with the permission of the auths. The tool should not
More informationConfounding by indication developments in matching, and instrumental variable methods. Richard Grieve London School of Hygiene and Tropical Medicine
Confounding by indication developments in matching, and instrumental variable methods Richard Grieve London School of Hygiene and Tropical Medicine 1 Outline 1. Causal inference and confounding 2. Genetic
More informationMethods for Addressing Selection Bias in Observational Studies
Methods for Addressing Selection Bias in Observational Studies Susan L. Ettner, Ph.D. Professor Division of General Internal Medicine and Health Services Research, UCLA What is Selection Bias? In the regression
More informationBiases in clinical research. Seungho Ryu, MD, PhD Kanguk Samsung Hospital, Sungkyunkwan University
Biases in clinical research Seungho Ryu, MD, PhD Kanguk Samsung Hospital, Sungkyunkwan University Learning objectives Describe the threats to causal inferences in clinical studies Understand the role of
More informationMidterm Exam ANSWERS Categorical Data Analysis, CHL5407H
Midterm Exam ANSWERS Categorical Data Analysis, CHL5407H 1. Data from a survey of women s attitudes towards mammography are provided in Table 1. Women were classified by their experience with mammography
More informationMarginal structural modeling in health services research
Technical Report No. 3 May 6, 2013 Marginal structural modeling in health services research Kyung Min Lee kml@bu.edu This paper was published in fulfillment of the requirements for PM931 Directed Study
More informationCan Quasi Experiments Yield Causal Inferences? Sample. Intervention 2/20/2012. Matthew L. Maciejewski, PhD Durham VA HSR&D and Duke University
Can Quasi Experiments Yield Causal Inferences? Matthew L. Maciejewski, PhD Durham VA HSR&D and Duke University Sample Study 1 Study 2 Year Age Race SES Health status Intervention Study 1 Study 2 Intervention
More informationSensitivity Analysis in Observational Research: Introducing the E-value
Sensitivity Analysis in Observational Research: Introducing the E-value Tyler J. VanderWeele Harvard T.H. Chan School of Public Health Departments of Epidemiology and Biostatistics 1 Plan of Presentation
More informationLecture Outline. Biost 590: Statistical Consulting. Stages of Scientific Studies. Scientific Method
Biost 590: Statistical Consulting Statistical Classification of Scientific Studies; Approach to Consulting Lecture Outline Statistical Classification of Scientific Studies Statistical Tasks Approach to
More informationImpact and adjustment of selection bias. in the assessment of measurement equivalence
Impact and adjustment of selection bias in the assessment of measurement equivalence Thomas Klausch, Joop Hox,& Barry Schouten Working Paper, Utrecht, December 2012 Corresponding author: Thomas Klausch,
More informationAcknowledgements. Section 1: The Science of Clinical Investigation. Scott Zeger. Marie Diener-West. ICTR Leadership / Team
INTRODUCTION TO CLINICAL RESEARCH Scientific Concepts for Clinical Research Karen Bandeen-Roche, Ph.D. July 15, 2013 Scott Zeger Acknowledgements Marie Diener-West ICTR Leadership / Team July 2013 JHU
More informationAdjusting overall survival for treatment switch
Adjusting overall survival for treatment switch Recommendations of a crossinstitutional statistical working group Claire Watkins Director and Consultant Statistician, Clarostat Consulting Ltd Cytel East
More informationEstimating the impact of community-level interventions: The SEARCH Trial and HIV Prevention in Sub-Saharan Africa
University of Massachusetts Amherst From the SelectedWorks of Laura B. Balzer June, 2012 Estimating the impact of community-level interventions: and HIV Prevention in Sub-Saharan Africa, University of
More informationChapter 21 Multilevel Propensity Score Methods for Estimating Causal Effects: A Latent Class Modeling Strategy
Chapter 21 Multilevel Propensity Score Methods for Estimating Causal Effects: A Latent Class Modeling Strategy Jee-Seon Kim and Peter M. Steiner Abstract Despite their appeal, randomized experiments cannot
More informationpharmacoepidemiology and drug safety 2016; 25(Suppl. 1): ORIGINAL REPORT
pharmacoepidemiology and drug safety 2016; 25(Suppl. 1): 114 121 Published online in Wiley Online Library (wileyonlinelibrary.com).3864 ORIGINAL REPORT Methodological comparison of marginal structural
More informationTuning Epidemiological Study Design Methods for Exploratory Data Analysis in Real World Data
Tuning Epidemiological Study Design Methods for Exploratory Data Analysis in Real World Data Andrew Bate Senior Director, Epidemiology Group Lead, Analytics 15th Annual Meeting of the International Society
More informationMS&E 226: Small Data
MS&E 226: Small Data Lecture 10: Introduction to inference (v2) Ramesh Johari ramesh.johari@stanford.edu 1 / 17 What is inference? 2 / 17 Where did our data come from? Recall our sample is: Y, the vector
More informationPropensity Score Methods for Analyzing Observational Data Like Randomized Experiments: Challenges and Solutions for Rare Outcomes and Exposures
American Journal of Epidemiology The Author 2015. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail:
More informationIntroduction. Step 2: Student Role - Your Plan of Action. Confounding. Good luck and have fun!
Confounding Introduction You have learned that random error and bias must be considered as possible explanations for an observed association between an exposure and disease. This week we will examine the
More informationAnalytic Strategies for the OAI Data
Analytic Strategies for the OAI Data Charles E. McCulloch, Division of Biostatistics, Dept of Epidemiology and Biostatistics, UCSF ACR October 2008 Outline 1. Introduction and examples. 2. General analysis
More informationHow to analyze correlated and longitudinal data?
How to analyze correlated and longitudinal data? Niloofar Ramezani, University of Northern Colorado, Greeley, Colorado ABSTRACT Longitudinal and correlated data are extensively used across disciplines
More informationCausal Mediation Analysis with the CAUSALMED Procedure
Paper SAS1991-2018 Causal Mediation Analysis with the CAUSALMED Procedure Yiu-Fai Yung, Michael Lamm, and Wei Zhang, SAS Institute Inc. Abstract Important policy and health care decisions often depend
More informationIMPROVING PROPENSITY SCORE METHODS IN PHARMACOEPIDEMIOLOGY
IMPROVING PROPENSITY SCORE METHODS IN PHARMACOEPIDEMIOLOGY Mohammed Sanni Ali Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, University of Utrecht,
More informationUsing directed acyclic graphs to guide analyses of neighbourhood health effects: an introduction
University of Michigan, Ann Arbor, Michigan, USA Correspondence to: Dr A V Diez Roux, Center for Social Epidemiology and Population Health, 3rd Floor SPH Tower, 109 Observatory St, Ann Arbor, MI 48109-2029,
More informationComparing treatments evaluated in studies forming disconnected networks of evidence: A review of methods
Comparing treatments evaluated in studies forming disconnected networks of evidence: A review of methods John W Stevens Reader in Decision Science University of Sheffield EFPSI European Statistical Meeting
More informationImplementing double-robust estimators of causal effects
The Stata Journal (2008) 8, Number 3, pp. 334 353 Implementing double-robust estimators of causal effects Richard Emsley Biostatistics, Health Methodology Research Group The University of Manchester, UK
More informationConfounding. Confounding and effect modification. Example (after Rothman, 1998) Beer and Rectal Ca. Confounding (after Rothman, 1998)
Confounding Confounding and effect modification Epidemiology 511 W. A. Kukull vember 23 2004 A function of the complex interrelationships between various exposures and disease. Occurs when the disease
More informationImputation approaches for potential outcomes in causal inference
Int. J. Epidemiol. Advance Access published July 25, 2015 International Journal of Epidemiology, 2015, 1 7 doi: 10.1093/ije/dyv135 Education Corner Education Corner Imputation approaches for potential
More informationComparison of propensity score methods and covariate adjustment:
Comparison of propensity score methods and covariate adjustment: Evaluation in four cardiovascular studies Markus C. Elze 1,2, John Gregson 1,Usman Baber 3, Elizabeth Williamson 1, Samantha Sartori 3,
More informationChapter 13 Estimating the Modified Odds Ratio
Chapter 13 Estimating the Modified Odds Ratio Modified odds ratio vis-à-vis modified mean difference To a large extent, this chapter replicates the content of Chapter 10 (Estimating the modified mean difference),
More informationThe Impact of Confounder Selection in Propensity. Scores for Rare Events Data - with Applications to. Birth Defects
The Impact of Confounder Selection in Propensity Scores for Rare Events Data - with Applications to arxiv:1702.07009v1 [stat.ap] 22 Feb 2017 Birth Defects Ronghui Xu 1,2, Jue Hou 2 and Christina D. Chambers
More informationGlossary From Running Randomized Evaluations: A Practical Guide, by Rachel Glennerster and Kudzai Takavarasha
Glossary From Running Randomized Evaluations: A Practical Guide, by Rachel Glennerster and Kudzai Takavarasha attrition: When data are missing because we are unable to measure the outcomes of some of the
More informationIntroduction to diagnostic accuracy meta-analysis. Yemisi Takwoingi October 2015
Introduction to diagnostic accuracy meta-analysis Yemisi Takwoingi October 2015 Learning objectives To appreciate the concept underlying DTA meta-analytic approaches To know the Moses-Littenberg SROC method
More informationPropensity Score Analysis Shenyang Guo, Ph.D.
Propensity Score Analysis Shenyang Guo, Ph.D. Upcoming Seminar: April 7-8, 2017, Philadelphia, Pennsylvania Propensity Score Analysis 1. Overview 1.1 Observational studies and challenges 1.2 Why and when
More informationn Outline final paper, add to outline as research progresses n Update literature review periodically (check citeseer)
Project Dilemmas How do I know when I m done? How do I know what I ve accomplished? clearly define focus/goal from beginning design a search method that handles plateaus improve some ML method s robustness
More informationStrategies for handling missing data in randomised trials
Strategies for handling missing data in randomised trials NIHR statistical meeting London, 13th February 2012 Ian White MRC Biostatistics Unit, Cambridge, UK Plan 1. Why do missing data matter? 2. Popular
More informationBiases in clinical research. Seungho Ryu, MD, PhD Kanguk Samsung Hospital, Sungkyunkwan University
Biases in clinical research Seungho Ryu, MD, PhD Kanguk Samsung Hospital, Sungkyunkwan University Learning objectives Describe the threats to causal inferences in clinical studies Understand the role of
More informationA COMPARISON OF IMPUTATION METHODS FOR MISSING DATA IN A MULTI-CENTER RANDOMIZED CLINICAL TRIAL: THE IMPACT STUDY
A COMPARISON OF IMPUTATION METHODS FOR MISSING DATA IN A MULTI-CENTER RANDOMIZED CLINICAL TRIAL: THE IMPACT STUDY Lingqi Tang 1, Thomas R. Belin 2, and Juwon Song 2 1 Center for Health Services Research,
More informationObjective: To describe a new approach to neighborhood effects studies based on residential mobility and demonstrate this approach in the context of
Objective: To describe a new approach to neighborhood effects studies based on residential mobility and demonstrate this approach in the context of neighborhood deprivation and preterm birth. Key Points:
More informationDaniel Boduszek University of Huddersfield
Daniel Boduszek University of Huddersfield d.boduszek@hud.ac.uk Introduction to Logistic Regression SPSS procedure of LR Interpretation of SPSS output Presenting results from LR Logistic regression is
More information